Taxis Toward Hydrogen Gas by Methanococcus maripaludis

Taxis Toward Hydrogen Gas byMethanococcus maripaludisKristen A. Brileya1,2,4*, James M. Connolly1,3, Carey Downey1,2, Robin Gerlach1,3 & Matthew W. Fields1,2,4

1Center for Biofilm Engineering, Montana State University, Bozeman, MT, USA, 2Department of Microbiology, Montana StateUniversity, Bozeman, MT, USA, 3Department of Chemical and Biological Engineering, Montana State University, Bozeman, MT,USA, 4ENIGMA (http://enigma.lbl.gov).

Knowledge of taxis (directed swimming) in the Archaea is currently expanding through identification ofnovel receptors, effectors, and proteins involved in signal transduction to the flagellar motor. Although theability for biological cells to sense and swim toward hydrogen gas has been hypothesized for many years, thiscapacity has yet to be observed and demonstrated. Here we show that the average swimming velocityincreases in the direction of a source of hydrogen gas for the methanogen, Methanococcus maripaludis usinga capillary assay with anoxic gas-phase control and time-lapse microscopy. The results indicate that amethanogen couples motility to hydrogen concentration sensing and is the first direct observation ofhydrogenotaxis in any domain of life. Hydrogenotaxis represents a strategy that would impart a competitiveadvantage to motile microorganisms that compete for hydrogen gas and would impact the C, S andN cycles.

Hydrogen gas (H2) is a crucial substrate for methanogens as well as a common source of energy for otherorganisms in both anaerobic and aerobic environments, including acetogens, sulfate- and sulfur-reducers,and hydrogen-oxidizers1–4. Biological methane (CH4) production from H2 and carbon dioxide (CO2)

contributes to greenhouse gas emissions and is possibly one of the oldest microbial metabolisms5,6. Under-standing the ecological strategies of methanogens is not only important for our knowledge of early earth processesand present-day anaerobic environments, but also for determining potential roles in human health conditions(e.g., colon cancer and periodontal disease), where positive correlations have been made with incidence of diseaseand occurrence of methanogens7,8. Methanococcus maripaludis is an anaerobic archaeum that can use H2 orformate as electron donor to reduce CO2 to CH4 and is considered a model mesophilic methanogen. Recently, theswimming behavior of M. maripaludis was described9, but chemotactic responses have not been shown.Chemotaxis has been demonstrated for Archaea, including methanogens10,11, but taxis to hydrogen has not beenshown for any domain of life. The chemotaxis signal transduction system in Archaea is similar to the well-studiedsystem in Bacteria; however, the flagellar switch is different and none of the archaeal flagellar proteins havehomologs to bacterial flagellar proteins12–15.

Chemotaxis has been the subject of many mathematical models and the majority have concentrated onreproducing the population-level observation of migrating bands of high cell concentration in swarm platesand capillary experiments16. Pioneering work in modeling chemotaxis behavior by Keller and Segel in 1971 hasbeen the basis of the most common mathematical models17. In one dimension, with x being the spatial variable,the Keller-Segel model can be described as a flux, J, such that

J~{mLbLx

zx(s)bLsLx

ð1Þ

where m is the cell diffusion coefficient that takes random, non-directed, movement of cells into account. b is themicrobial population density, s is the attractant concentration and x(s) is the non-constant chemotactic coef-ficient. The population flux, J, can be differentiated to yield the more common form

LbLt

~{LJLx

~{LLx

{mLbLx

zx(s)bLsLx

� �ð2Þ

and the average cell swimming velocity, v, is calculated by dividing the flux by the population density, or v 5 J/b.Lapidus and Schiller18 proposed a form of x(s) such that

OPEN

SUBJECT AREAS:ENVIRONMENTAL

MICROBIOLOGY

ARCHAEA BIOLOGY

Received1 July 2013

Accepted18 October 2013

Published5 November 2013

Correspondence andrequests for materials

should be addressed toM.W.F. (matthew.

[email protected])

*Current address:Department of Biology,

Portland StateUniversity, Portland,

OR, USA

SCIENTIFIC REPORTS | 3 : 3140 | DOI: 10.1038/srep03140 1

x(s)~xkd

(kdzs)2 ð3Þ

where x is the constant chemotactic coefficient and kd is the receptor-ligand binding dissociation constant. The Lapidus-Schiller x(s) term,and variations thereof, have been used widely to describe chemotaxisin bacteria19. Most work on archaeal chemotaxis has been performedwith Halobacterium salinarum; however, attempts to mathemat-ically describe population flux in archaea have not been published.In addition, x and kd have not been determined for any archaea12,20–22.

The goal of the present study was to subject M. maripaludis cells toa H2 concentration gradient and compare swimming behavior tomodel predictions. Attractant (H2) transport was modeled byFickian diffusion and consumption by the population was modeledin the Michaelis–Menten form such that

LsLt

~DL2sLx2

{rmaxbs

(kmzs)ð4Þ

where D is the diffusion coefficient for H2, rmax is the maximumconsumption rate, and km is the half-saturation constant.

A modified capillary assay was used in which cells were loaded intoa gas-tight anaerobic capillary under an anoxic atmosphere, and avalve allowed controlled exposure to a H2 source (see SupplementaryFig. S1 and S2 online). In previous experiments, capillaries that con-tained a dissolved chemoattractant were immerged into a cell sus-pension and cells entered the capillary in the presence of achemoattractant over a known incubation time. The populationincrease in the capillary relative to a control was then quantifiedthrough cell enumeration23. For this study, a novel method wasdeveloped to directly observe microscopic swimming behavior ofanaerobic cells inside a capillary during exposure to a gas. M. mar-ipaludis cells were tracked for direction and velocity changes uponexposure to H2 or an Argon (Ar) control to quantify the averagepopulation-wide response. Cell movement was measured usingtime-lapse confocal laser scanning microscopy (CLSM) in the centerof the capillary (0.5 cm from the gas phase; as in Supplementary Fig.S1 online). A reaction-diffusion model predicted that within 10–15minutes H2 concentrations would reach the threshold of 2.5–23 mM,at which hydrogenotrophic methanogens have been shown to use H2

in pure culture studies24–27 (Figure 1).

ResultsWhen M. maripaludis cells were exposed to H2, the average swim-ming velocity exhibited significant bias toward H2 within ten min-utes (Figure 2A). Biased random walk was not observed when cellswere exposed to an Ar gradient (Figure 2B), nor was swimmingvelocity affected normal to the H2 or Ar gradients (Figures 2C andD). A strong chemotactic response was observed when cells were firststarved of H2 for 4–5 hours, while cultures that had not been starveddid not show an increase in biased swimming (see SupplementaryFig. S3 online). Similar results were observed in Rhodobacter sphaer-oides and Sinorhizobium meliloti when well-fed cells exhibitedweaker chemotactic responses than starved cells toward attractantssuch as organic acids and sugars. It was suggested that this wasbecause chemosensory pathways in these organisms were dependenton metabolic state and/or transport28, as could be the case for H2

sensing in M. maripaludis.The starvation period (4–5 h) that was required to induce a tract-

able response to H2 did not cause loss of motility, as observed duringstarvation of H. salinarum20. The average swimming speed of starvedM. maripaludis cells was 2.1 6 0.05 mm s21 before exposure to H2

(Figure 3A) and increased to 3.1 6 0.02 mm s21 after exposure; equalto non-starved cells. Maximum swimming speed averages (calcu-lated by averaging the single maximum speed from each time point)were highly variable between time points, and the average maximum

was higher for non-starved cells (9.3 mm s21 versus 8.5 mm s21 forstarved cells 1 H2) and lowest for starved cells before H2 exposure(7.0 mm s21) (Figure 3B). The highest maximum observed speed was91 mm s21 for H2 starved cells after exposure to H2 (Figure 3B innerboxes), approximately twice the previously observed maximumspeed of 45 mm s21 for M. maripaludis9.

A Keller-Segel model was applied with boundary conditions spe-cific to our study in an effort to quantify the observed chemotacticresponse to H2 in context with other known chemotactic responses.Currently, K-S model parameter values do not exist for any organismin Archaea or for any other gas besides O2; therefore, a broad range ofkd values were used in the model. Three unknowns, namely ligand-receptor dissociation constant (kd), chemotactic coefficient (x), andrandom cell diffusion coefficient (m), were independently fitted byvarying one unknown across the range of published literature values,while keeping the other two variables constant at the average lite-rature values (see Supplementary Table S1 online). The best fitfor average swimming velocity was obtained with a kd value of0.70 mM, with 0.30 and 2.30 mM corresponding to the 95% confid-ence interval of the data, whilex and mwere kept at average publishedvalues (Figure 4A). This range for kd is similar to that observed forEscherichia coli AW405 to a-methyl aspartate29. It is; however, quitedifferent from values reported for Bacillus subtilis receptor affinity toO2 (0.0015 and 0.075 mM for the high and low affinity of the recep-tor, respectively), which is the only previously reported kd for a gas30.

The model could also be fitted to the velocity curve by varying xand keeping kd and m constant at average literature values(Figure 4B). The average x of 9.3 3 1023 cm2 s21 used to fit theexperimental results is higher than the literature range 7.20 3 1025

to 1.24 3 1023 cm2 s21 (see Supplementary Table S1 online). The

Figure 1 | The predicted hydrogen concentration over time at theobservation point 0.5 cm from the gas phase over the course of theexperiment shown with (A) linear axes and (B) log-linear axes. The 2.5–

23 mM threshold, at which hydrogenotrophic methanogens have been

shown to use H2 in pure culture studies24–27, is reached at approximately 10

minutes.

www.nature.com/scientificreports


experimental results were best fitted to the model that assumed theaverage published value for x, and then by varying kd (Figure 4A).The model could not be fitted to the experimental data by varying maccording to published values (Figure 4C, note the difference in y-axis scale); however, changing m allowed for a change in shape of thevelocity curve.

The response time predicted by the model does not fit the rapidresponse observed in the experiment and this is likely due to aninaccurate prediction of mass transport inside the capillary (for bothH2 and cells). Both random cell diffusion (m) and the diffusion ofhydrogen would require correction from predicted values if masstransport was inaccurately represented. First, considering m indepen-dently, it is likely that response time would be faster for cells with alarger m (Figure 4C). This is based on the proportionality of m toswimming speed and run time, and inverse proportionality to one

minus the cosine of the turn angle h31. Although not measureddirectly here, M. maripaludis has been shown to have relatively longruns and small changes in direction or turn angles9. Turn anglesbetween 20 and 45u would result in the largest value of m for a givenswimming speed and run time, so it is reasonable to assume m wouldbe higher for this type of swimming. The observed rapid response canbe explained by an inaccurate prediction of H2 mass flux into theliquid domain. If H2 were to reach the observation point faster thanpredicted by the model, then one would predict a proportionallyfaster response. This was an entirely static system on a visuallyobservable scale but it is possible that unpredictable micro-scaleconvective forces increased H2 mass transport. The collective,directed motion of swimming cells may have induced convectiveflow inside the capillary. Previous work estimates that a force of0.5 pN is exerted by a cell swimming at 25 mm s21 32. While

Figure 2 | Average cell velocity (black) with 95% confidence intervals (gray) before and after opening valve to gas phase for (A) H2 and (B) Ar control.Positive y-axis values indicate movement toward the gas phase, negative y-axis values indicate movement away from the gas phase. Average cell velocity

(C) and (D) normal to concentration gradient (n 5 5 for H2 and n 5 3 for Ar).

Figure 3 | Swimming speed of M. maripaludis cells when starved for 5 hours before H2 exposure (2H2 and n 5 4,725), after H2 exposure (1H2 and n 520,189), and when not starved (No starve, n 5 45,001). Values represent the mean and error bars represent 95% confidence intervals. (A) Average

swimming speed. The difference between –H2 and both 1H2 and no starvation is significant (p , 0.05). (B) Average maximum swimming speeds are

represented by bars, and absolute maximum values inside each bar represent the highest observed speed over all time points for the condition. Differences

between all conditions are significant (p , 0.05) and p values were calculated with two-tailed t-tests.



chemotaxis-induced convective flow is difficult to demonstrate con-clusively, increased mass transport in the capillary could affect masstransport predictions.

A diffusion coefficient correction term, d, was introduced as a toolto investigate the effect of increased mass transport on the predictedresponse time. The correction was applied to the H2 diffusion coef-ficient and the random cell diffusion coefficient (m) simultaneously,making the assumption that micro-scale convection was the primarysource of error. Any convection would affect the random diffusion ofcells and the diffusion of H2 equally. Increased mass transport in theliquid (simulating convection) caused an earlier response in all con-ditions that were investigated (see Supplementary Fig. S4 online).The value of d is likely to be on the order of 1 3 1029 to 1 3

1028 cm2 s21 based on curve shape and response time as comparedto the experimental velocity data.

Another possible explanation for the more rapid experimentalresponse is the presence of more than one type of H2 receptor withvarying affinities. This would allow M. maripaludis to respond to H2

across a wider range of concentrations. The predicted kd value0.7 mM is high, and the presence of a high affinity receptor in M.maripaludis is likely. B. subtilis O2 receptors with two distinct affin-ities and binding components have been shown30. The presented data

demonstrate a chemotactic response to H2, but more focused physio-logical studies and subsequent model refinement are needed to betterrepresent this poorly understood phenomenon.

In our capillary assay, cells were not observed to accumulate inbands despite the chemotactic response observed. Typical chemo-tactic bands observed in capillary and swarm plate assays result fromcell accumulation as a net result of a biased random walk16. In thecapillary assay used, convection inside the capillary may have pre-vented cell accumulation. It is also possible that there are factorsinvolved in chemotactic band formation that are unique betweenorganisms or factors specific to hydrogenotaxis (e.g., swimmingmode, adaption response, quorum sensing).

In our work, banding-like behavior was observed on a larger scalewhen M. maripaludis batch cultures were grown statically with H2.Under these conditions a pellicle formed at the gas-liquid interface(Figure 5A, B). This is a similar observation to the one made byBeijernck in 1893, where aerotactic cells were observed swimmingtoward the meniscus of a test tube16. When cultures of M. maripa-ludis were grown statically with the soluble electron donor formate,cells grew throughout the liquid medium (not in a pellicle) and hadless cell-associated carbohydrate than H2-grown pellicle cultures(see Supplementary Fig. S5 online). These results suggest that

Figure 4 | Model fitting results (smooth black lines) overlaid on experimental results with 95% confidence intervals (grey lines) (A) using averageliterature values for x and m and varying kd over the range shown or (B) using average literature values for kd and m while varying x over the range shown(C) or m varied across the range of literature values.



extra-polymeric substance (EPS) production is important for theformation of a pellicle in response to a H2 gradient and may requirea time-scale longer than used in the capillary experiment.

DiscussionThe hydrogenotrophic methanogen, M. maripaludis, displays che-motactic behavior toward H2 gas (hydrogenotaxis). Changes in pHwere not observed in medium with or without H2 under the testedgrowth conditions and times, and these results indicate that theobserved taxis was in response to H2 and not H1. Although thecommonly used Keller-Segel model was not able to predict the exactresponse with previously published parameter ranges (values), para-meter adjustment allowed the model to replicate the trend on a scalesimilar to other organisms and chemoattractants. The ability to movetoward higher concentrations of H2 could incur an advantage tomethanogens that are otherwise outcompeted by those that are ableto use H2 at lower concentrations and/or utilize terminal electronacceptors that are more energetically favorable. The demonstratedchemotactic response would also allow cells to maintain desirablelocalization with respect to the major energy source as well as allowfor proximity to H2-producers in mixed communities. Thus, hydro-genotaxis could play a crucial role in the establishment andmaintenance of microbial interactions at the population- and com-munity-level. The observed hydrogenotaxis could represent a wide-spread eco-physiological strategy of methanogens and otherhydrogen-utilizing microbes that are important to processes suchas bioremediation and overall carbon cycling4,33. CH4 is a potentgreenhouse gas that has an estimated global warming potential(GWP) 25–40% higher than CO2 per molecule34. The three largestcontributions of CH4 to atmospheric flux as of 2010 (wetlands,ruminant emissions, and rice cultivation) are the net result of theactivity of anaerobic communities dominated by the exchange ofH2

35,36. To the best of our knowledge this is the first direct observationof hydrogenotaxis in any domain of life.

MethodsCulturing conditions. Methanococcus maripaludis strain S2 was grown in Balchtubes or serum bottles fitted with black butyl stoppers (Geo-Microbial TechnologiesInc., Ochelata, OK) and aluminum crimp seals. Methanococcus Culture Medium(MCC) was prepared under a stream of anoxic 80% N2, 20% CO2 and contains perliter 0.33 g KCl, 2.7 g, MgCl2.6H2O, 3.5 g MgSO4N7H2O, 0.14 g CaCl2N2H2O, 0.5 gNH4Cl, 5 g NaHCO3, 22 g NaCl, 0.14 g K2HPO4, 5 mL FeSO4 solution (0.19 gFeSO4N7H2O/100 mL of 10 mM HCl), 1 mL trace metal solution (per 100 mL; 2.1 gNa3CitrateN2H2O, adjust pH to 6.5, then 0.45 g MnSO4NH2O, 0.1 g CoCl2N6H2O,0.1 g ZnSO4N7H2O, 0.01 g CuSO4N5H2O, 0.01 g AlK(SO4)2, 0.01 g H3BO4, 0.1 gNa2MoO4N2H2O, 0.025 g NiCl2N6H2O, 0.2 g Na2SeO3, 0.01 g VCl3, 0.0033 gNa2WO4N2H2O) 10 mL of vitamin solution (per liter; 2 mg biotin, 2 mg folic acid,10 mg pyridoxine HCl, 5 mg thiamine HCl, 5 mg riboflavin, 5 mg nicotinic acid,5 mg DL-calcium pantothenate, 0.1 mg vitamin B12, 5 mg r-aminobenzoinc acid,5 mg lipoic acid), 1 mL of Resazurin solution (1 g/L). This solution is boiled under astream of gas before adding 0.5 g cysteineNH2O.HCl, then cooled under gas37. Thesolution was dispensed anaerobically and autoclaved. After inoculation, theheadspace was displaced and pressurized to 25 PSI with anoxic 80% H2:20% CO2

through a sterile filter. Modified MCC medium was used for growth experiments with

formate, where NaCl was reduced to 10.5 g/L; 200 mM Na-formate and 200 mM 3-(N-morpholino) propanesulfonic acid (MOPS) were added, and no H2 was added.

Cultures for taxis experiments were grown at 30uC (with no shaking) from frozenglycerol stocks in 40 mL MCC in 125 mL serum bottles to late exponential phase,then transferred to 5 mL MCC (10% inoculum) in 18 3 150 mm Balch tubes. Theseworking cultures were grown to stationary phase (around 120 h and average celldensity 69,420 cells/mL). 1 mL of culture was added to 5 mL fresh MCC in a Balchtube with no H2 and either incubated for 4–5 h at 30uC under starvation conditions(no electron source), or used immediately for non-starvation conditions. Media saltswere separated from cell suspension prior to taxis experiments by centrifuging theinverted Balch tube (100 3 g; 30uC for 10 minutes).

Electron microscopy. Pellicles were removed from tubes with a plastic inoculatingloop and placed on a round coverslip freshly treated with poly-L-Lysine (1 mg/mL)and fixed in a solution of 2% paraformaldehyde, 2.5% glutaraldehyde and 0.05 M Na-cacodylate overnight at room temperature. Coverslips were rinsed and stepwisedehydrated in ethanol, and then critical point dried on a Samdri-795 (tousimis,Rockville, MD). Coverslips were mounted on SEM stubs with double-sided carbontape and colloidal silver, and then sputter coated with Iridium for 35 s at 35 mA.Images were collected on a Zeiss Supra55VP FE-SEM.

Capillary assay. Square glass capillary tubes (1.0 mm) with the ends fitted withnorprene tubing connected to a polypropylene female luer-lock hose barb adapter(Cole-Parmer, Vernon Hills, IL) were partially filled with the cell suspension in ananaerobic chamber that contained only N2 and CO2. A 5 mL glass gas-tight luer locksyringe (SGE, Inc., Austin, TX) was used to transfer the cells to the capillary, and leftattached to the capillary with the valve closed. A second 5 mL glass gas-tight syringewith 100% H2 was attached to the gas side of the capillary (see Supplementary Fig. S1online). The entire syringe/capillary assembly was removed from the anaerobicchamber and placed in a petri dish water bath on the microscope stage and firmlysecured with poster putty and tape (see Supplementary Fig. S3 online).

Microscopic observation of swimming behavior and image analysis. A Leica TCSSP5 II upright confocal microscope was enclosed in an incubation chamber and wasequilibrated to 30uC. High-resolution time lapse images were collected every 0.753 sat the center of the capillary, 0.5 cm from the cell suspension/gas phase interface. A 253 water-dipping objective was used and a 3 3 optical zoom was applied resulting in afinal field of view of 206.9 3 206.9 mm2 and a pixel size of 0.20 mm. Images wereinitially acquired for 10 minutes with N2/CO2 in the gas phase, then the valve wasopened to the H2 syringe and images were captured for 40 minutes. The controlexperiments were identical to the above except that 100% Ar was used instead of H2,and images were acquired for a minimum of 8 minutes before the valve was openedand 38 minutes after.

Images were manually thresholded and binarized using MetaMorph v. 7.6(Molecular Devices, Sunnyvale, CA). Binary images were analyzed using Imaris v.7.5.2 (Bitplane, Inc., South Windsor, CT) with a particle-tracking module (ImarisTrack). 1 s, 5 s, and 10 s filter durations were tested where a given track was onlyanalyzed if it was as long or longer than the specified duration. The 5 s filter was usedfor the described analyses, and differences were not observed in overall trendsbetween track lengths.

Chemotaxis model. A one dimensional finite element model was constructed usingComsol Multiphysics Version 4.3 a that solves Equations 2 and 4 simultaneously inthe liquid domain. Diffusion of hydrogen through the gas domains was modeled byMaxwell-Stefan equations. All model parameters were corrected for temperature(30uC) and salinity (2.65% m/v) of the medium, where possible, and hydrogenconsumption rate parameters were estimated from literature values. SupplementaryTable S1 online shows all constants used in the model.

The geometry of the model consisted of one liquid domain and two gas domainsseparated by a valve that opens at t 5 0 to start the diffusion of hydrogen into thesystem. The short segment between the valve and the far right boundary is the lengthof the connection between the valve and the main volume of the gas-tight syringe. Thegeometry for each experimental replicate was slightly different so average lengthswere used. The diffusion of hydrogen through the gas domain of the capillary was

Figure 5 | (A) and (B) Field emission scanning electron micrographs of M. maripaludis pellicle with extracellular material and cells (indicated byarrows in (A)).



expected to be much faster than diffusion through the liquid domain so the preciselength of the gas domain was not expected to affect the analysis. The diffusion-limiting segment between the liquid-gas interface and the point of observation 0.5 cminto the liquid remained constant through all replicates and is therefore not anaverage value. Initial conditions are zero hydrogen inside the capillary, 100%hydrogen behind the valve in the second gas domain and a constant populationdensity in the liquid domain. Supplementary Figure S1 online shows the geometry,initial conditions and boundary conditions used in the model.

The model geometry was meshed at a maximum element size of 0.1 mm and timestep discretization was done with a backward differentiation formula (BDF) method.The time steps taken by the solver were allowed to be free with larger steps being takenat later time points where gradients are shallower. A relative solution tolerance of10220 was imposed and PARDISO38 was the general direct solver as implemented inComsol.

Most parameters in the model are well known or able to be calculated from theliterature with the exception of the chemotaxis constants x, m and kd. A literaturesearch was performed for observed chemotaxis values in any organism to establish arange of reasonable values. Twelve applicable values were found for x, twenty-one form and seven for kd. Maximum, average and minimum values for each are found inTable S1. The model was fitted to the average swimming velocity data by indepen-dently varying one chemotaxis parameter by trial and error while keeping the othertwo constant at the average literature value found in Supplementary Table S1 online.

Every effort was made to model the diffusion of hydrogen into the liquid domainaccurately; however, the likelihood of unpredictable factors such as micro-scalemixing due to convection still exists. To provide the flexibility to correct for enhancedmass transport, a correction term was applied such that

D0Haq~DHaqzd ð5Þ

Where d represents the mass transport enhancement beyond what is predicted fromFickian diffusion alone and D’Haq is the corrected diffusion coefficient used in themodel for this analysis. Similarly, d was also applied to the random cell diffusioncoefficient such that

m0~mzd ð6Þ

because any correction applied to DHaq would need to be applied to m on grounds thatconvection would affect the movement of cells the same as the pass transport of H2.

Carbohydrate and protein measurements. Protein concentrations were determinedwith the Lowry assay using bovine serum albumin as the standard39. Hexose sugarswere measured by the colorimetric cysteine-sulfuric acid method with glucose as thestandard. Pentose sugars were measured with a colorimetric orcinol-FeCl3 assay withxylose as the standard. A colorimetric carbazole assay was used to measure uronicacid concentration with D-galacturonic acid as the standard40.

1. Ferrera, I. et al. Diversity of 16S rRNA gene, ITS region and aclB gene of theAquificales. Extremophiles 11, 57–64 (2006).

2. Reysenbach, A. L., Longnecker, K. & Kirshtein, J. Novel Bacterial and ArchaealLineages from an In Situ Growth Chamber Deployed at a Mid-Atlantic RidgeHydrothermal Vent. Applied and Environmental Microbiology 66, 3798–3806(2000).

3. Poehlein, A. et al. An Ancient Pathway Combining Carbon Dioxide Fixation withthe Generation and Utilization of a Sodium Ion Gradient for ATP Synthesis. PLoSONE 7, 1–8 (2012).

4. Petersen, J. M. et al. Hydrogen is an energy source for hydrothermal ventsymbioses. Nature 476, 176–180 (2012).

5. Thauer, R. K. The Wolfe cycle comes full circle. Proc. Natl. Acad. Sci. U.S.A. 109,15084–15085 (2012).

6. Ueno, Y., Yamada, K., Yoshida, N., Maruyama, S. & Isozaki, Y. Evidence fromfluid inclusions for microbial methanogenesis in the early Archaean era. Nature440, 516–519 (2006).

7. Jarrell, K. F. et al. Major players on the microbial stage: why archaea are important.Microbiology 157, 919–936 (2011).

8. Conway de Macario, E. & Macario, A. J. L. Methanogenic archaea in health anddisease: A novel paradigm of microbial pathogenesis. Int. J. Med. Microbiol. 299,99–108 (2009).

9. Herzog, B. & Wirth, R. Swimming Behavior of Selected Species of Archaea.Applied and Environmental Microbiology 78, 1670–1674 (2012).

10. Migas, J., Anderson, K. L., Cruden, D. L. & Markovetz, A. J. Chemotaxis inMethanospirillum hungatei. Applied and Environmental Microbiology 55,264–265 (1989).

11. Sment, K. A. & Konisky, J. Chemotaxis in the archaebacterium Methanococcusvoltae. Journal of Bacteriology 171, 2870–2872 (1989).

12. Schlesner, M. et al. Identification of Archaea-specific chemotaxis proteins whichinteract with the flagellar apparatus. BMC Microbiol 9, 56–56 (2009).

13. Lacal, J., Garcıa-Fontana, C., Munoz-Martınez, F., Ramos, J.-L. & Krell, T. Sensingof environmental signals: classification of chemoreceptors according to the size oftheir ligand binding regions. Environ Microbiol 12, 2873–2884 (2010).

14. Thomas, N. A. & Jarrell, K. F. Characterization of Flagellum Gene Families ofMethanogenic Archaea and Localization of Novel Flagellum Accessory Proteins.Journal of Bacteriology 183, 7154–7164 (2001).

15. Porter, S. L., Wadhams, G. H. & Armitage, J. P. Signal processing in complexchemotaxis pathways. Nat Rev Micro 9, 153–165 (2011).

16. Adler, J. Chemotaxis in Bacteria. Science 153, 708–716 (1966).17. Keller, E. F. & Segel, L. A. Model for chemotaxis. Journal of Theoretical Biology 30,

225–234 (1971).18. Lapidus, I. R. & Schiller, R. Model for the chemotactic response of a bacterial

population. Biophysj 16, 779–789 (1976).19. Tindall, M. J., Maini, P. K., Porter, S. L. & Armitage, J. P. Overview of

Mathematical Approaches Used to Model Bacterial Chemotaxis II: BacterialPopulations. Bull. Math. Biol. 70, 1570–1607 (2008).

20. Streif, S., Staudinger, W. F., Marwan, W. & Oesterhelt, D. Flagellar Rotation in theArchaeon Halobacterium salinarum Depends on ATP. Journal of MolecularBiology 384, 1–8 (2008).

21. del Rosario, R. C. et al. Modelling the CheY D10K, Yl00W Halobacterium salinarummutant: sensitivity analysis allows choice of parameter to be modified in thephototaxis model. IET systems biology 1, 207–221 (2007).

22. Alatyrev, A. G. et al. Identification of the New Protein Participitating in theArchaea Motility Regulation. Biol Membrany 27, 121–132 (2010).

23. Adler, J. & Dahl, M. M. A Method for Measuring the Motility of Bacteria and forComparing Random and Non-random Motility. J Gen Microbiol 46, 161–173(1967).

24. Thauer, R. K., Kaster, A.-K., Seedorf, H., Buckel, W. & Hedderich, R.Methanogenic archaea: ecologically relevant differences in energy conservation.Nat Rev Micro 6, 579–591 (2008).

25. Lovley, D. R. Minimum threshold for hydrogen metabolism in methanogenicbacteria. Applied and Environmental Microbiology 49, 1530–1531 (1985).

26. Kaster, A.-K., Moll, J., Parey, K. & Thauer, R. K. Coupling of ferredoxin andheterodisulfide reduction via electron bifurcation in hydrogenotrophicmethanogenic archaea. Proc. Natl. Acad. Sci. U.S.A. 108, 2981–2986 (2011).

27. Ver Eecke, H. C. et al. Hydrogen-limited growth of hyperthermophilicmethanogens at deep-sea hydrothermal vents. Proc. Natl. Acad. Sci. U.S.A. 109,13674–13679 (2012).

28. Armitage, J. P. & Schmitt, R. Bacterial chemotaxis: Rhodobacter sphaeroides andSinorhizobium meliloti-variations on a theme? Microbiology 143, 3671–3682(1997).

29. Ahmed, T. & Stocker, R. Experimental Verification of the Behavioral Foundationof Bacterial Transport Parameters Using Microfluidics. Biophysical Journal 95,4481–4493 (2008).

30. Zhang, W., Olson, J. S. & Phillips Jr, G. N. Biophysical and KineticCharacterization of HemAT, an Aerotaxis Receptor from Bacillus subtilis.Biophysj 88, 2801–2814 (2005).

31. Lewus, P. & Ford, R. M. Quantification of Random Motility and ChemotaxisBacterial Transport Coefficients Using Individual-Cell and Population-ScaleAssays. Biotechnol. Bioeng. 75, 292–304 (2001).

32. Kim, M. & Kim, T. Diffusion-Based and Long-Range Concentration Gradients ofMultiple Chemicals for Bacterial Chemotaxis Assays. Anal. Chem. 82, 9401–9409(2010).

33. Kleiner, M. et al. Metaproteomics of a gutless marine worm and its symbioticmicrobial community reveal unusual pathways for carbon and energy use. Proc.Natl. Acad. Sci. U.S.A. 109, E1173–E1182 (2012).

34. Shindell, D. T. et al. Improved Attribution of Climate Forcing to Emissions.Science 326, 716–718 (2009).

35. Neef, L., van Weele, M. & van Velthoven, P. Optimal estimation of the present-dayglobal methane budget. Global Biogeochem. Cycles 24, 1–10 (2010).

36. Schlesinger, W. H. & Bernhardt, E. S. Biogeochemistry: An Analysis of GlobalChange. 3rd Ed. (Academic Press, Waltham, MA, 2013).

37. Miller, T. L. & Wolin, M. J. A Serum Bottle Modification of the HungateTechnique for Cultivating Obligate Anaerobes. Appl Microbiol 27, 985–987(1974).

38. Schenk, O. & Gartner, K. Solving unsymmetric sparse systems of linear equationswith PARDISO. Future Gener Comput Syst 20, 475–487 (2004).

39. Lowry, O. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. Protein Measurementwith the Folin Phenol Reagent. J biol chem 193, 265–275 (1951).

40. Chaplin, M. F. & Kennedy, J. F. Carbohydrate Analysis: A Practical Approach. ThePractical Approach Series. 2nd Ed. (Irl Press, Oxford, 1994).

AcknowledgmentsThe authors wish to thank Sara Altenburg for helping with the pH experiment, Betsey Pittsfor microscopy assistance, Gill Geesey for encouraging us to pursue the experiments, AlParker for assistance with statistics, Adam Arkin and Roland Hatzenpichler for helpfulcomments, and William B. Whitman for his suggestion to use formate. This workconducted by ENIGMA- Ecosystems and Networks Integrated with Genes and MolecularAssemblies (http://enigma.lbl.gov), a Scientific Focus Area Program at Lawrence BerkeleyNational Laboratory, was supported by the Office of Science, Office of Biological andEnvironmental Research, of the U. S. Department of Energy under Contract No.DE-AC02-05CH11231. K.A.B. and J.M.C. were also supported by a NSF-IGERT fellowshipin Geobiological Systems at Montana State University (DGE 0654336). Partial support forR.G. was provided through the National Science Foundation under CHE-1230632. Theconfocal microscopy equipment used was purchased with funding from the NSF-MajorResearch Instrumentation Program and the M.J. Murdock Charitable Trust.



Author contributionsK.A.B. developed experimental design, performed experiments, critically evaluated themodel, wrote and revised the manuscript. J.M.C. developed experimental design, performedexperiments, created the model, wrote and revised the manuscript. C.D. performedexperiments and revised the manuscript. R.G. developed experimental design, criticallyevaluated the model, and revised the manuscript. M.W.F. developed experimental design,critically evaluated the model, and revised the manuscript.

Additional informationSupplementary information accompanies this paper at http://www.nature.com/scientificreports

Competing financial interests: The authors declare no competing financial interests.

How to cite this article: Brileya, K.A., Connolly, J.M., Downey, C., Gerlach, R. & Fields,M.W. Taxis Toward Hydrogen Gas by Methanococcus maripaludis. Sci. Rep. 3, 3140;DOI:10.1038/srep03140 (2013).

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported license. To view a copy of this license,

visit http://creativecommons.org/licenses/by-nc-sa/3.0



http://www.nature.com/scientificreports

http://www.nature.com/scientificreports

http://creativecommons.org/licenses/by-nc-sa/3.0

Taxis Toward Hydrogen Gas by Methanococcus maripaludis

Documents